
1 étoile --------- Chaque utilisateur a rédigé en moyenne 1.4 reviews Chaque établissement a reçu en moyenne 12.3 reviews
Langues détectées : ----------------- en :: 10000 occurrences
Détermination du score de cohérence pour un nombre déterminé de sujets, entre 2 et 30
<AxesSubplot:xlabel='nb_topics', ylabel='coherence_score'>
Liste des mots les plus fréquents pour un regroupement LDA en 21 sujets :
(20, '0.031*"bar" + 0.014*"bartender" + 0.013*"friend" + 0.012*"like" + 0.011*"drink"') (6, '0.017*"said" + 0.014*"back" + 0.014*"told" + 0.014*"would" + 0.013*"manager"') (7, '0.017*"steak" + 0.015*"ordered" + 0.013*"meal" + 0.011*"restaurant" + 0.011*"meat"') (13, '0.022*"restaurant" + 0.016*"menu" + 0.015*"price" + 0.015*"reservation" + 0.013*"would"') (16, '0.017*"like" + 0.009*"one" + 0.009*"people" + 0.008*"get" + 0.007*"want"') (8, '0.084*"room" + 0.042*"hotel" + 0.022*"car" + 0.017*"stay" + 0.017*"casino"') (19, '0.060*"order" + 0.060*"pizza" + 0.017*"ordered" + 0.017*"time" + 0.013*"delivery"') (18, '0.031*"cheese" + 0.029*"salad" + 0.018*"bread" + 0.017*"sauce" + 0.013*"like"') (14, '0.015*"rice" + 0.014*"sushi" + 0.014*"chicken" + 0.014*"fish" + 0.013*"roll"') (9, '0.014*"chicken" + 0.013*"like" + 0.012*"sandwich" + 0.012*"fry" + 0.012*"order"') (0, '0.075*"taco" + 0.040*"mexican" + 0.036*"chip" + 0.027*"salsa" + 0.027*"donut"') (2, '0.015*"like" + 0.014*"better" + 0.013*"good" + 0.010*"taste" + 0.009*"price"') (12, '0.040*"husband" + 0.028*"vegan" + 0.020*"hair" + 0.010*"law" + 0.009*"hookah"') (3, '0.035*"owner" + 0.030*"service" + 0.030*"customer" + 0.030*"review" + 0.029*"restaurant"') (17, '0.015*"dirty" + 0.012*"health" + 0.011*"clean" + 0.011*"one" + 0.009*"hand"') (10, '0.018*"drink" + 0.017*"table" + 0.013*"server" + 0.010*"one" + 0.010*"asked"') (1, '0.054*"cake" + 0.021*"mother" + 0.015*"mom" + 0.014*"birthday" + 0.013*"child"') (15, '0.039*"ramen" + 0.031*"bug" + 0.030*"pork" + 0.029*"roach" + 0.028*"com"') (11, '0.020*"coffee" + 0.019*"time" + 0.017*"like" + 0.016*"get" + 0.010*"really"') (4, '0.232*"burger" + 0.066*"dog" + 0.030*"store" + 0.027*"bun" + 0.026*"fry"')
Représentation graphique des bigrammes pour $K=21$
C:\Users\aledo\anaconda3\envs\p6\lib\site-packages\pyLDAvis\_prepare.py:247: FutureWarning: In a future version of pandas all arguments of DataFrame.drop except for the argument 'labels' will be keyword-only.
by='saliency', ascending=False).head(R).drop('saliency', 1)
Liste des mots les plus fréquents pour un regroupement LDA en 21 sujets :
(6, '0.023*"owner" + 0.012*"customer" + 0.011*"restaurant" + 0.010*"business" + 0.009*"people"') (15, '0.052*"steak" + 0.018*"pasta" + 0.012*"meal" + 0.011*"bread" + 0.007*"well_done"') (8, '0.046*"table" + 0.021*"restaurant" + 0.014*"minute" + 0.014*"food" + 0.014*"server"') (16, '0.030*"groupon" + 0.013*"cupcake" + 0.012*"fianc" + 0.008*"told" + 0.008*"voucher"') (14, '0.055*"order" + 0.017*"food" + 0.016*"time" + 0.015*"get" + 0.014*"place"') (1, '0.049*"shrimp" + 0.037*"salad" + 0.028*"crab" + 0.018*"seafood" + 0.016*"ramen"') (12, '0.013*"would" + 0.011*"manager" + 0.009*"said" + 0.008*"told" + 0.008*"back"') (17, '0.033*"room" + 0.019*"hotel" + 0.012*"one" + 0.011*"night" + 0.010*"stay"') (13, '0.051*"bar" + 0.043*"drink" + 0.021*"bartender" + 0.019*"beer" + 0.019*"friend"') (10, '0.013*"place" + 0.009*"like" + 0.008*"restaurant" + 0.008*"one" + 0.006*"also"') (5, '0.018*"food" + 0.012*"place" + 0.011*"one" + 0.009*"good" + 0.009*"like"') (20, '0.026*"food" + 0.022*"chicken" + 0.017*"place" + 0.014*"dish" + 0.013*"sauce"') (2, '0.165*"pizza" + 0.013*"slice" + 0.012*"cheese" + 0.011*"pie" + 0.010*"crust"') (19, '0.075*"cake" + 0.022*"bagel" + 0.022*"donut" + 0.019*"bakery" + 0.013*"mask"') (18, '0.016*"place" + 0.012*"food" + 0.011*"like" + 0.009*"good" + 0.008*"one"') (3, '0.014*"food" + 0.012*"one" + 0.009*"order" + 0.008*"time" + 0.008*"place"') (4, '0.091*"buffet" + 0.013*"dirty" + 0.012*"store" + 0.010*"mom" + 0.008*"went_counter"') (7, '0.017*"coupon" + 0.010*"club" + 0.009*"cat" + 0.008*"racist" + 0.008*"place"') (0, '0.018*"dog" + 0.010*"card" + 0.009*"service_dog" + 0.008*"receipt" + 0.006*"food_cold"') (9, '0.047*"taco" + 0.019*"burrito" + 0.011*"tortilla" + 0.011*"crawfish" + 0.009*"oyster"')
Représentation graphique des bigrammes pour $K=21$
C:\Users\aledo\anaconda3\envs\p6\lib\site-packages\pyLDAvis\_prepare.py:247: FutureWarning: In a future version of pandas all arguments of DataFrame.drop except for the argument 'labels' will be keyword-only.
by='saliency', ascending=False).head(R).drop('saliency', 1)
Topic 0: said | manager | asked | told | back | never | came | would | went | could Topic 1: salad | ordered | sauce | steak | dish | shrimp | meal | soup | like | menu Topic 2: pizza | crust | slice | topping | cheese | delivery | pepperoni | pie | oven | sauce Topic 3: table | minute | waitress | wait | seated | waited | hostess | waiting | came | sat Topic 4: bar | drink | bartender | beer | friend | night | glass | bouncer | one | group Topic 5: order | minute | delivery | time | hour | phone | called | call | ordered | get Topic 6: food | good | cold | mexican | eat | quality | price | got | buffet | hour Topic 7: reservation | party | made | group | told | would | hostess | hour | dinner | groupon Topic 8: burger | fry | medium | bun | cheese | patty | shake | ordered | beer | onion Topic 9: cake | wedding | bakery | birthday | cupcake | day | crab | would | chocolate | icing Topic 10: coffee | breakfast | egg | cup | shop | toast | counter | morning | barista | starbucks Topic 11: sushi | roll | fish | chef | rice | eat | japanese | nigiri | tuna | piece Topic 12: room | hotel | stay | desk | bed | casino | night | front | floor | day Topic 13: restaurant | review | experience | star | yelp | would | one | menu | dining | could Topic 14: service | customer | rude | horrible | terrible | never | worst | bad | staff | ever Topic 15: sandwich | cheese | bread | fry | meat | lunch | cheesesteak | cuban | philly | onion Topic 16: chicken | fried | wing | rice | sauce | piece | raw | taco | flavor | tasted Topic 17: place | like | time | get | people | really | even | good | one | say Topic 18: card | credit | tip | bill | charge | cash | charged | pay | receipt | check Topic 19: owner | business | customer | employee | mask | rude | establishment | support | treat | review Topic 20: server | drink | came | check | water | took | asked | come | brought | never
Topic 0: came back | drink order | another minute | waited minute | took minute | take order | waited waited | minute get | never came | dining experience Topic 1: customer service | poor customer | worst customer | terrible customer | horrible customer | service never | service skill | service ever | bad customer | good customer Topic 2: first time | last time | even though | last night | year old | look like | behind counter | looked like | food good | somewhere else Topic 3: could give | zero star | star would | give zero | give star | wish could | negative star | would give | give place | star could Topic 4: credit card | card company | card number | gift card | day later | card machine | gave credit | next day | card charge | left cash Topic 5: come back | never come | food cold | bed bug | ever come | back place | would come | dim sum | back say | told come Topic 6: happy hour | hour menu | hour special | hour price | said happy | sit bar | hour bar | come happy | behind bar | went friday Topic 7: tasted like | taste like | fried rice | indian food | like came | looked like | mashed potato | sauce tasted | chicken tasted | like sitting Topic 8: bad service | bad food | food bad | service horrible | service bad | better option | management completely | horrible management | food service | back bad Topic 9: new orleans | red bean | restaurant new | bean rice | place new | service new | ever new | thought tourist | orleans would | kitchen nightmare Topic 10: one star | star review | give one | gave one | star rating | place one | star food | five star | get one | giving one Topic 11: food poisoning | got food | get food | two word | word food | hour later | case one | poisoning eat | star case | got sick Topic 12: would never | never back | recommend place | place anyone | never recommend | would recommend | unprofessional would | horrible service | food subpar | place would Topic 13: gluten free | free pizza | free item | free option | money back | free menu | ordered gluten | behind counter | celiac disease | free bread Topic 14: going back | never going | definitely going | ever going | needle say | put bar | beer expensive | say going | eat dinner | back anytime Topic 15: wearing mask | wear mask | without mask | one wearing | health department | employee wearing | staff wearing | social distancing | mask optional | client base Topic 16: waste time | time money | much better | save time | money place | complete waste | better waste | meat food | non smoking | service better Topic 17: service dog | service animal | veteran service | american disability | disability act | federal law | veteran day | mediterranean cuisine | dog restaurant | pizza place Topic 18: minute later | five minute | came back | back minute | another minute | order minute | ten minute | later still | later came | hour later Topic 19: mexican food | mexican restaurant | taco bell | real mexican | authentic mexican | chip salsa | food ever | worst mexican | good mexican | food eat Topic 20: mac cheese | fried chicken | green bean | soul food | taste like | potato salad | chicken wing | pulled pork | collard green | sweetie pie
Word most similar to : wait ['waiting', 'waited', 'sit', 'took', 'ready', 'later', 'closing', 'take', 'seated', 'timed']
Word most similar to : poisoning ['diarrhea', 'symptom', 'sick', 'ache', 'stomach', 'coli', 'tainted', 'ill', 'shipment', 'salmonella']
Word most similar to : service ['experience', 'attitude', 'food', 'treatment', 'inattentive', 'consistently', 'ambiance', 'overall', 'exceptionally', 'atmosphere']
Word most similar to : food ['meal', 'sushi', 'everything', 'service', 'pizza', 'entree', 'drink', 'apps', 'dish', 'actually']
Word most similar to : drink ['beer', 'cocktail', 'water', 'refill', 'beverage', 'round', 'soda', 'bartender', 'mimosa', 'appetizer']
Word most similar to : restaurant ['place', 'establishment', 'eatery', 'location', 'joint', 'considering', 'resturant', 'dining', 'people', 'venue']
Word most similar to : taste ['flavor', 'tasted', 'aftertaste', 'texture', 'consistency', 'marinade', 'tasteless', 'bland', 'overpowering', 'flavorless']
Word most similar to : accessibility ['braille', 'ramp', 'handicapped', 'accessible', 'height', 'congested', 'leading', 'wooden', 'lighting', 'loo']
Word most similar to : dirty ['sticky', 'grimy', 'wipe', 'clean', 'wiped', 'filthy', 'silverware', 'cloth', 'mop', 'dingy']
Word most similar to : racism ['prejudice', 'discrimination', 'discriminating', 'undertone', 'bigoted', 'negativity', 'condone', 'cruelty', 'irresponsible', 'wikipedia']
Word most similar to : waiter ['waitress', 'server', 'hostess', 'table', 'bartender', 'manager', 'waitstaff', 'maitre', 'sever', 'host']
Word most similar to : waitress ['waiter', 'server', 'bartender', 'lady', 'manager', 'table', 'hostess', 'girl', 'finally', 'waitstaff']
Word most similar to : bartender ['server', 'waitress', 'drink', 'waiter', 'bar', 'female', 'guy', 'girl', 'tab', 'bouncer']
Word most similar to : price ['pricing', 'priced', 'quality', 'cost', 'expectation', 'expensive', 'amount', 'dollar', 'portion', 'carte']
Word most similar to : quality ['average', 'priced', 'subpar', 'mediocre', 'turnover', 'price', 'substandard', 'execution', 'superb', 'inconsistent']
Word most similar to : review ['rating', 'reviewer', 'yelpers', 'yelp', 'post', 'yelper', 'feedback', 'bash', 'experience', 'update']
Word most similar to : star ['rating', 'negative', 'rate', 'scathing', 'zagat', 'stellar', 'coveted', 'bash', 'merit', 'chance']
Word most similar to : like ['stupid', 'embarassed', 'strange', 'think', 'kind', 'blankly', 'badly', 'funny', 'crap', 'shit']
Word most similar to : order ['minute', 'min', 'ordering', 'ordered', 'check', 'forever', 'told', 'payment', 'pick', 'ask']
Word most similar to : table ['seat', 'booth', 'seated', 'sat', 'party', 'server', 'waitress', 'hostess', 'plate', 'waiter']
Word most similar to : bar ['upstairs', 'bartender', 'patio', 'table', 'downstairs', 'booth', 'lounge', 'bench', 'outside', 'watched']
Word most similar to : owner ['manager', 'manger', 'management', 'supervisor', 'employee', 'swearing', 'kevin', 'pamela', 'verbatim', 'behalf']
Word most similar to : minute ['min', 'hour', 'patiently', 'order', 'minuet', 'moment', 'ten', 'forever', 'finally', 'eternity']
Word most similar to : customer ['patron', 'costumer', 'employee', 'business', 'client', 'job', 'management', 'people', 'respect', 'profession']
Word most similar to : reservation ['appointment', 'arrangement', 'party', 'booking', 'seated', 'reserved', 'advance', 'hostess', 'beforehand', 'opentable']
Word most similar to : experience ['service', 'visit', 'review', 'treatment', 'impression', 'gamlin', 'incident', 'evening', 'experienced', 'encounter']
Catégories : ---------- 1 : inside 2 : outside 3 : drink 4 : food 5 : menu
<matplotlib.image.AxesImage at 0x1d8033b3ee0>
Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).
<matplotlib.image.AxesImage at 0x1d80341ca00>
inside -------
outside -------
menu -------
food -------
drink -------
Score ARI après t-SNE : 0.1175142007713266
Répartition des photos dans chacune des 5 classes trouvées par le K-Means :
<AxesSubplot:>
Matrice de confusion :
array([[ 61, 69, 27, 28, 15],
[ 32, 96, 42, 22, 8],
[ 49, 20, 84, 39, 8],
[ 13, 9, 21, 45, 112],
[ 63, 19, 81, 21, 16]], dtype=int64)
Rapport de classification :
precision recall f1-score support
0 0.28 0.30 0.29 200
1 0.45 0.48 0.46 200
2 0.33 0.42 0.37 200
3 0.29 0.23 0.25 200
4 0.10 0.08 0.09 200
accuracy 0.30 1000
macro avg 0.29 0.30 0.29 1000
weighted avg 0.29 0.30 0.29 1000
Model: "vgg16"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_6 (InputLayer) [(None, 224, 224, 3)] 0
block1_conv1 (Conv2D) (None, 224, 224, 64) 1792
block1_conv2 (Conv2D) (None, 224, 224, 64) 36928
block1_pool (MaxPooling2D) (None, 112, 112, 64) 0
block2_conv1 (Conv2D) (None, 112, 112, 128) 73856
block2_conv2 (Conv2D) (None, 112, 112, 128) 147584
block2_pool (MaxPooling2D) (None, 56, 56, 128) 0
block3_conv1 (Conv2D) (None, 56, 56, 256) 295168
block3_conv2 (Conv2D) (None, 56, 56, 256) 590080
block3_conv3 (Conv2D) (None, 56, 56, 256) 590080
block3_pool (MaxPooling2D) (None, 28, 28, 256) 0
block4_conv1 (Conv2D) (None, 28, 28, 512) 1180160
block4_conv2 (Conv2D) (None, 28, 28, 512) 2359808
block4_conv3 (Conv2D) (None, 28, 28, 512) 2359808
block4_pool (MaxPooling2D) (None, 14, 14, 512) 0
block5_conv1 (Conv2D) (None, 14, 14, 512) 2359808
block5_conv2 (Conv2D) (None, 14, 14, 512) 2359808
block5_conv3 (Conv2D) (None, 14, 14, 512) 2359808
block5_pool (MaxPooling2D) (None, 7, 7, 512) 0
global_max_pooling2d_2 (Glo (None, 512) 0
balMaxPooling2D)
=================================================================
Total params: 14,714,688
Trainable params: 0
Non-trainable params: 14,714,688
_________________________________________________________________
C:\Users\aledo\anaconda3\envs\p6\lib\site-packages\sklearn\cluster\_kmeans.py:1332: UserWarning: KMeans is known to have a memory leak on Windows with MKL, when there are less chunks than available threads. You can avoid it by setting the environment variable OMP_NUM_THREADS=4. warnings.warn(
Score ARI brut : 0.4598008242990195
Score ARI après ACP : 0.471538173585247
C:\Users\aledo\anaconda3\envs\p6\lib\site-packages\sklearn\cluster\_kmeans.py:1332: UserWarning: KMeans is known to have a memory leak on Windows with MKL, when there are less chunks than available threads. You can avoid it by setting the environment variable OMP_NUM_THREADS=4. warnings.warn(
Score ARI après t-SNE : 0.533229525218234
Répartition des photos dans chacune des 5 classes trouvées par le K-Means :
<AxesSubplot:>
Matrice de confusion :
Correspondance des clusters : [4, 3, 1, 2, 0] [[149 18 1 29 3] [ 5 178 4 9 4] [ 3 2 156 31 8] [ 85 11 6 97 1] [ 0 3 18 9 170]]
Rapport de classification :
precision recall f1-score support
0 0.62 0.74 0.67 200
1 0.84 0.89 0.86 200
2 0.84 0.78 0.81 200
3 0.55 0.48 0.52 200
4 0.91 0.85 0.88 200
accuracy 0.75 1000
macro avg 0.75 0.75 0.75 1000
weighted avg 0.75 0.75 0.75 1000
C:\Users\aledo\anaconda3\envs\p6\lib\site-packages\sklearn\cluster\_kmeans.py:1332: UserWarning: KMeans is known to have a memory leak on Windows with MKL, when there are less chunks than available threads. You can avoid it by setting the environment variable OMP_NUM_THREADS=4. warnings.warn(
Score ARI brut : 0.536333646979273
Score ARI après ACP : 0.5434840625386322
C:\Users\aledo\anaconda3\envs\p6\lib\site-packages\sklearn\cluster\_kmeans.py:1332: UserWarning: KMeans is known to have a memory leak on Windows with MKL, when there are less chunks than available threads. You can avoid it by setting the environment variable OMP_NUM_THREADS=4. warnings.warn(
Score ARI après t-SNE : 0.589022343063851
Répartition des photos dans chacune des 5 classes trouvées par le K-Means :
<AxesSubplot:>
Matrice de confusion :
Correspondance des clusters : [0, 2, 4, 3, 1] [[146 7 5 42 0] [ 3 186 4 4 3] [ 2 2 173 20 3] [ 85 10 7 95 3] [ 0 3 12 10 175]]
Rapport de classification :
precision recall f1-score support
0 0.62 0.73 0.67 200
1 0.89 0.93 0.91 200
2 0.86 0.86 0.86 200
3 0.56 0.47 0.51 200
4 0.95 0.88 0.91 200
accuracy 0.78 1000
macro avg 0.78 0.78 0.77 1000
weighted avg 0.78 0.78 0.77 1000
| Métrique | VGG16 | ResNet50 |
|---|---|---|
| ARI brute | 0.46 | 0.53 |
| ARI après ACP | 0.47 | 0.54 |
| ARI après t-SNE | 0.53 | 0.59 |
| Acuracy | 0.75 | 0.78 |
Récupération de 593 nouveaux avis à propos de 200 restaurants. Requête effectuée en 14.4 secondes.
| business_id | stars | text | |
|---|---|---|---|
| 0 | Vaq49W0ubGjuIc4h5_qQ0w | 5 | This place was absolutely delicious! It was a ... |
| 1 | Vaq49W0ubGjuIc4h5_qQ0w | 4 | The food here is spot on. Portions were very w... |
| 2 | Vaq49W0ubGjuIc4h5_qQ0w | 4 | This place is very popular for their beautiful... |
| 3 | oK_SLmmAVQg3meguh7LrIA | 5 | Fantastic chicken, beef, and mango lasses. Col... |
| 4 | oK_SLmmAVQg3meguh7LrIA | 5 | Just wow. Can't go wrong here. I'm Indian I wo... |